Mode of action of modified and unmodified ...

Biochimie 84 (2002) 545–557

Mode of action of modified and unmodified bacteriocins from Gram-positive bacteria Yann Héchard a,*, Hans-Georg Sahl b a

Laboratory of Fundamental and Applied Microbiology, University of Poitiers, 40, avenue du Recteur-Pineau, 86022 Poitiers, cedex, France b Institute for Medical Microbiology and Immunology, University of Bonn, Sigmund Freud Strasse 25, 53127 Bonn, Germany Received 6 March 2002; accepted 7 June 2002

Abstract The antibiotic activity of bacteriocins from Gram-positive bacteria, whether they are modified (class I bacteriocins, lantibiotics) or unmodified (class II), is based on interaction with the bacterial membrane. However, recent work has demonstrated that for many bacteriocins, generalised membrane disruption models as elaborated for amphiphilic peptides (e.g. tyriodal pore or carpet model) cannot adequately describe the bactericidal action. Rather, specific targets seem to be involved in pore formation and other activities. For the nisin and epidermin family of lantibiotics, the membrane-bound cell wall precursor lipid II has recently been identified as target. The duramycin family of lantibiotics binds specifically to phosphoethanolamine which results in inhibition of phospholipase A2 and various other cellular functions. Most of the class II bacteriocins dissipate the proton motive force (PMF) of the target cell, via pore formation. The subclass IIa bacteriocin activity likely depends on a mannose permease of the phosphotransferase system (PTS) as specific target. The subclass IIb bacteriocins (two-component) also induce dissipation of the PMF by forming cation- or anion-specific pores; specific targets have not yet been identified. Finally, the subclass IIc comprises miscellaneous peptides with various modes of action such as membrane permeabilisation, specific inhibition of septum formation and pheromone activity. © 2002 Société française de biochimie et biologie moléculaire / Éditions scientifiques et médicales Elsevier SAS. All rights reserved. Keywords: Bacteriocin; Lantibiotic; Antibacterial; Pore.

1. Introduction Antibacterial peptides or proteins ribosomally synthesised by bacteria are commonly referred to as bacteriocins. Most of those produced by Gram-positive bacteria are synthesised by lactic acid bacteria (for a review see [1,2]). The bacteriocins are generally secreted via a dedicated ABC transporter but some of them are secreted via a secdependent pathway. The bacteriocin gene is usually associated to a gene encoding the so-called immunity protein. The latter protects bacteria from their own bacteriocin, although the protection mechanism remains still unclear for most immunity proteins. Bacteriocins classification has been Abbreviations: Dha, didehydroalanine; Dhb, didehydrobutyrine; MIC, minimum inhibitory concentration; PMF, proton motive force; CF, carboxyfluoresceine; PTS, phosphotransferase system * Corresponding author. Tel.: +33-0-549-454-007; fax: +33-0-549-453503. E-mail address: [email protected] (Y. Héchard).

proposed on the basis of their primary structure [1,2]. The class I is composed by modified peptides, named lantibiotics. The class II gathers unmodified peptides. The class III is composed by large bacteriocins whose mode of action was poorly studied. Thus, the latter will not be described in this review.

2. Modified bacteriocins (class I) A substantial proportion of the peptide bacteriocins of Gram-positive bacteria undergo extensive post-translational modification before they are exported from the cell (class I bacteriocins). The most prominent modifications include the dehydration of serine and threonine residues to didehydroalanine (Dha) and didehydrobutyrine (Dhb), respectively, and the subsequent addition of suitably located cysteine residues via their SH groups to the C=C double bond of Dha or Dhb. The thioether amino acids resulting

© 2002 Société française de biochimie et biologie moléculaire / Éditions scientifiques et médicales Elsevier SAS. All rights reserved. PII: S 0 3 0 0 - 9 0 8 4 ( 0 2 ) 0 1 4 1 7 - 7

546

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557

from this addition reaction are, in contrast to disulfide bridges, acid stable and can be identified in peptide hydrolysates as lanthionine (originating from Dha and Cys) and 3-methyllanthionine (from Dhb and Cys). It was the presence of the lanthionines that led to proposing the designation ‘lantibiotics’ as an abbreviation for lanthioninecontaining peptide antibiotics. Details on the chemistry of lantibiotics and the biosynthesis pathways can be found in several recent reviews [3–6]. It should be noted that, although the unusual residues may give the impression that lantibiotics are related to classical peptide antibiotics such as gramicidin or valinomycin, the ribosomal biosynthesis clearly classify them as bacteriocins [4,7]. It appears that only the presence of dedicated enzymes for amino acid dehydration and thioether formation can distinguish biosynthetic gene clusters of lantibiotics from those of class II bacteriocins [4]. Ever since the structure of the most prominent lantibiotic nisin (Fig. 1) had been elucidated [8] it was speculated that the unusual amino acids may be important for function. Particularly the dehydro residues were discussed to represent ideal sites for addition of thiol groups and other nucleophiles. Such covalent modifications were thought to result in inactivation of enzymes, coenzymes and other compounds with functionally important groups. Indeed, there is evidence that such a mechanism is involved in the sporicidal activity of nisin and subtilin [9]. However, it appeared unlikely that such a specialised mechanism could account for the rather generalised broad-spectrum antibiotic

activity of nisin and related lantibiotics such as epidermin, subtilin, Pep5, mersacidin and others. The first report on the mode of action of nisin appeared as early as 1953 by Ramseier [10] who described a detergent-like membrane disruption based on the observation that nisin treated bacteria rapidly leaked UV-absorbing cellular material such as nucleotides. Later studies unravelled an inhibitory role of nisin in cell wall biosynthesis [11] which could then be attributed to interaction of the lantibiotic with the membrane-bound cell wall precursors lipid I and lipid II (Fig. 2) [12]. However, subsequent investigations with growing bacteria clearly showed that a great majority of the cells were killed rapidly along with an instant depolarisation of the cytoplasmic membrane, leakage of cellular material and complete stop of biosynthetic processes [13–15]. The apparent discrepancies in the interpretation of the molecular nature of the nisin killing activity were only recently solved and united into one mode-ofaction model [16–18]. It was demonstrated that nisin and epidermin use lipid II primarily as a docking molecule for formation of highly effective, targeted pores. Simultaneously, cell wall biosynthesis is inhibited through arresting the precursor in the membrane. In addition, nisin is able to form non-targeted pores as proposed earlier [19] and, particularly in staphylococci, it activates cell wall hydrolysing enzymes [20,21]. Thus, its antibiotic activity is based on a multiplicity of activities which may combine differently for individual target bacteria and explain the range of sensitivities of various bacterial species. In the following,

Fig. 1. Structure of the lantibiotics nisin (a) and mersacidin (b); residues which derive from posttranslational modification of Ser, Thr and Cys are marked in gray; Dha, dehydroalanine; Dhb, dehydrobutyrine; Abu, aminobutyric acid; Ala–S–Ala, lanthionine; Abu–S–Ala, 3-methyllanthionine; for further details on the amino acid modifications see text and [3–7].

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557

Fig. 2. Structure of the ultimate membrane-bound cell wall precursor undecaprenylpyrophosphoryl-MurNAc(pentapeptide)-GlcNAc, lipid II.

the individual activities of lantibiotics are described in more detail. 2.1. Amphiphilic cationic lantibiotics (type-A) can form target-independent pores Cationic amphiphilic peptides are wide-spread in nature. Throughout the living world, they are produced by microorganisms to antagonise competitors and by plants and insects and vertebrates as effector molecules to prevent and combat microbial infections. The destructive effects of amphiphilic compounds on membranes with a negative surface charge has been known for a long time and there is a wealth of information on the biophysical behaviour of synthetic and natural peptides [22]. In line with the early observation of Ramseier [10], it was initially assumed that nisin and related lantibiotics similarly disintegrate microbial membranes. There are numerous reports demonstrating that these peptides kill bacterial cells by interference with energy transduction occurring at the cytoplasmic membrane. Addition of nisin and other type-A peptides immediately inhibits biosynthesis processes of macromolecules such as DNA, RNA, proteins and polysaccharides [15]. Furthermore, bacterial cells were unable to actively take-up amino acids and become leaky for inorganic ions and small metabolites [13]. A concept of energisation-dependent activity was deduced from experiments with cytoplasmic vesicles and intact cells as well as with model membranes. Conductance measurements using artificial bilayer membranes (black lipid membranes, BLM) were in good agreement with the results obtained with intact cells and physiological membranes [14,23]. There was no macroscopic membrane conductivity below a certain threshold potential; the minimum voltage ranged from –50 to –100 mV for various peptides, and calculated pore diameters were between 1 nm for nisin and about 2 nm for subtilin with a lifetime in the millisecond range [23–25]. Initially, it was discussed that a ‘barrel-stave’ model may adequately describe the activity of these lantibiotics. A barrel-stave model, in which α-helical amphipathic peptides

547

bind via electrostatical interaction to the outer leaflet in a parallel orientation to the membrane surface, had been elaborated e.g. by Boheim [21] and others [22] for alamethicin. In order to avoid the unfavourable position of polar residues to the lipid acyl chains several monomers of peptides have to assemble forming a bundle of helices. After insertion of the peptides into the membrane the non-polar side chains of the peptides interact with the hydrophobic lipid core of the membrane and the hydrophilic side chains point inward, which results in the formation of a water-filled pore. Both, the size and the stability of the barrel-stave channel, depend on the number of peptides involved in pore formation. Based on conformational data derived from NMR studies with nisin in the presence of membrane-mimicking micelles [26,27] a ‘wedge model’ for pore formation by lantibiotics was proposed (Fig. 3a) [27,28]. This model takes into account that type-A lantibiotics are rather flexible in aqueous solution and defined structural elements have only been identified in small thioether rings [29,30]. Upon contact with the membrane, the peptides adopt an amphiphilic conformation with the charged residues aligned to one face of the molecules and the hydrophobic residues aligned to the other. The cationic peptides interact with the phospholipid head groups by ionic forces causing a locally disturbed bilayer structure, while the hydrophobic residues insert into the membrane. Studies with nisin demonstrated that the C-terminal region as well as the overall negative surface charge of the membrane [31–35] were important for binding and pore formation. It is assumed that several molecules have to associate within the membrane in order to form a pore, since lantibiotics are too short to transverse the membrane more than once. On the other hand, several transmembrane segments are necessary to form a pore with a diameter of 1 nm or more as deduced from BML studies. The formation of pores leads to dissipation of the membrane potential and promotes a rapid efflux of small metabolites such as amino acids or ATP, which in turn immediately stops all cellular biosynthetic processes as mentioned above. 2.2. Nisin and epidermin: formation of high affınity targetmediated pores The above model may describe the behaviour of peptides in pure lipid bilayers where high peptide concentrations are needed to induce effects; also, the killing activity for those microbes which do have micromolar minimum inhibitory concentrations (MICs) may well be primarily based on such a mechanism. However, nisin and related bacteriocins frequently kill bacteria in the nanomolar concentration range, indicating that additional activities or specific targets may be involved. Since type-A lantibiotics can act on artificial membranes, a concept of specific targets being involved in the membrane interaction had not been considered before. However, for nisin, a finite number of binding

548

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557

a

b

c

Fig. 3. Molecular activities of of type-A and B lantibiotics. (a) at micromolar concentrations, the cationic type-A peptides (nisin, epidermin, Pep5 and others) from wedge-like, target-independent pores. (b) at nanomolar concentrations, nisin and epidermin form target-mediated pores using lipid II as a docking molecule. (c) mersacidin and actagardine also bind to lipid II but only block its incorporation into peptidoglycan; such an activity is intrinsic to nisin and epidermin as well, and becomes obvious with mutant peptides which are deficient in pore formation [18].

sites and specific antagonisation of nisin activity by the inactive N-terminal nisin fragment 1–12 had been observed [36], indicating that a defined binding site may be blocked by the fragment. With that respect, it was important to recall publications by Linnett and Strominger [11] who reported that nisin inhibits peptidoglycan biosynthesis, and that it binds to the membrane-bound peptidoglycan precursor undecaprenylpyrophosphoryl-MurNAc(pentapeptide)GlcNAc, the so-called lipid II. These reports raised the idea that lipid II may be involved on the formation of pores. This was subsequently shown for

nisin and epidermin using lipid II supplemented liposomes [16]. Detailed in vitro studies then demonstrated that lipid II serves as a docking molecule for specific binding to the bacterial membrane (Fig. 3b) [17,18]. Genetically engineered nisin variants helped to identify the structural requirements for the interaction of the peptide with lipid II [18]. Mutations affecting the conformation of the N-terminal part of nisin comprising rings A through C (e.g. [S3T]nisin), led to reduced binding and increased the peptide concentration necessary for pore formation; i.e. the binding constant for the [S3T]mutant was 0.043 × 107 (M–1)

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557

as compared to 2 × 107 (M–1) for the wild type peptide and the minimum concentration for pore formation increased from the 2 to 50 nM range. Peptides with mutations in the flexible hinge region (e.g. [∆N20/∆M21]nisin) were completely inactive in the pore formation assay. When tested in vivo against living bacteria, their activity was only reduced to some extent. The remaining in vivo activity was shown to be due to the unaltered capacity of the mutated peptide to bind to lipid II, and thus to inhibit the incorporation of lipid II into the peptidoglycan network. The N-terminal part of nisin is essential for binding to the membrane bound cell wall precursor lipid II and the resulting inhibition of the peptidoglycan synthesis. In contrast to the nisin induced pore formation without a docking-molecule, a negative surface charge of the membrane is not necessary. However, the positively charged C-terminus of nisin is still important for the initial binding to the anionic cell wall polymer and thus for antimicrobial activity in vivo. Since the N-terminal double ring system of nisin is instrumental for the interaction with lipid II, it seems likely that subtilin and the entire group of epidermin-like lantibiotics [3,4,7] act similarly. All these lantibiotics share the general setup of this ring system and, in fact, epidermin itself has been shown to inhibit peptidoglycan synthesis and to promote liposome leakage [16]. The combination of two killing mechanisms, inhibition of the peptidoglycan synthesis and pore formation, in one molecule obviously potentiates in vivo the antibiotic activity and results in nanomolar MIC values, a strategy that may well be worth considered for the construction of novel antibiotics. Details on the molecular architecture of the lipid II-mediated pores remain to be elucidated, although it seems likely that several such stochiometric peptide: lipid II complexes must assemble to form a pore. 2.3. Mersacidin, actagardine, nisin and epidermin: inhibition of peptidoglycan biosynthesis via complex formation with lipid II The exploration of the above described duality of nisin and epidermin activity was greatly influenced by preceding studies on the type-B lantibiotics mersacidin and actagardine. Both are active against a variety of Gram-positive bacteria, with actagardine being most effective against streptococci and obligate anaerobes [37], while mersacidin (Fig. 2) is almost equally active against staphylococci, streptococci, bacilli, clostridia, corynebacteria, peptostreptococci, and Propionibacterium acnes [38,39]. Gramnegative bacteria are not susceptible, since peptides cannot pass the outer membrane of bacteria. Although both lantibiotics in vitro have only moderate MIC values, mersacidin has attracted recent attention due to its significant in vivo activity. It effectively cured systemic staphylococcal infections in mice including those caused by methicillin-resistant

549

Staphylococcus aureus (MRSA strains), as well as subcutaneous staphylococcal abscesses in rats. In both cases the lantibiotic equalled or even exceeded the activity of vancomycin [39]. Mersacidin and actagardine were shown to interfere with the cell wall biosynthesis by inhibiting the incorporation of glucose and D-alanine into cell wall material of Staphylococcus simulans 22, whereas DNA, RNA and protein synthesis proceeded unhindered. Both peptides inhibit peptidoglycan biosynthesis at the level of transglycosylation by forming a complex with the membrane-bound peptidoglycan precursor lipid II [40,41]. The binding of [14C]mersacidin to growing cells as well as to isolated membranes capable of in vitro peptidoglycan synthesis was strictly dependent on the availability of lipid II, and antibiotic inhibitors of lipid II formation strongly interfered with binding of mersacidin. Furthermore, labelled mersacidin tightly associated with micelles formed from purified and labelled lipid II, and the addition of isolated lipid II to the culture broth efficiently antagonised the bactericidal activity of mersacidin. The molecular target site of mersacidin and actagardine on lipid II differs from that of nisin (see above) and of the glycopeptide antibiotic vancomycin, which binds to lipid II via the C-terminal D-Ala–D-Ala of the pentapeptide side-chain; in contrast, mersacidin and actagardine rather interact with the disaccharide-pyrophosphate moiety of lipid II (Fig. 3c) [41]. The two lantibiotics contain one ring structure that has been almost completely conserved in both molecules, indicating its importance for activity [40,42]. 2.4. Additional activities of type-A lantibiotics In addition to pore formation and inhibition of the cell wall biosynthesis, nisin and the related cationic lantibiotic Pep5 have been shown to induce autolysis of susceptible staphylococcal cells, resulting in massive cell wall degradation, most markedly in the area of the septa between dividing daughter cells. The peptides are able to release two cell wall hydrolysing enzymes, an N-acetylmuramoyl-Lalanine amidase and an N-acetylglucosaminidase, which are strongly cationic proteins binding to the cell wall via electrostatic interactions with the negatively charged teichoic-, teichuronic-, and lipoteichoic acids; tight binding to these polymers keeps the autolysins inactive. The cationic peptides displace enzymes from the cell wall intrinsic inhibitors by a cation exchange-like process, resulting in apparent enzyme activation and rapid cell lysis [20,21]. Furthermore, nisin and subtilin inhibit the germination of bacterial spores. This activity obviously depends on the presence of Dha residues in position 5 of both peptides [9]. It is presumed that the double bond provides a reactive group for interaction with a spore-associated factor that is essential for outgrowth of spores.

550

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557

2.5. Cinnamycin-like lantibiotics: complex formation with phosphatidylethanol-amine The cinnamycin-like type-B lantibiotics display antibactericidal activity against a few bacterial strains, in particular specific Bacillus strains. Treated cells show increased membrane permeability [43] as well as an impaired ATPdependent protein translocation [44] and calcium uptake [45]. In addition, duramycin and cinnamycin inhibit a number of metabolic processes in eucaryotic cells [46,47]; e.g. induction of hemolysis of erythrocytes [48] or inhibition of phospholipase A2, thereby interfering with prostaglandin and leucotriene biosynthesis [49]. These biological activities had usually been identified individually, e.g. in pharmaceutical screening assays and little effort was undertaken to deduce a common motif for molecular details. However, it is reasonable to assume that such activities may be explained on the molecular level by the ability of this group of lantibiotics to form a specific complex with phosphatidylethanol-amine [49]. 2.6. Conclusions As a general scheme it can be stated that lantibiotics act on the cytoplasmic membrane of bacteria. Some, e.g. mersacidin, actagardine and the cinnamycin group, bind to specific lipid or lipid-bound targets and inhibit subsequent enzyme reactions. The cationic amphiphilic lantibiotics, at micromolar concentrations, disrupt the membrane in a non-targeted fashion. Additionally, they may also bind to specific targets and use it for high-affinity, target-mediated pore formation. For nisin and the epidermin group, the target has been identified and is obviously the same as for the non-pore forming mersacidin and actagardine, i.e. lipid

II. For others such as Pep5, target-mediated poration is likely, although the docking molecule remains to be identified. Interestingly, when a specific target is used for pore formation, it is simultaneously blocked from its natural function, thus yielding very effective combined antibiotic activities by the corresponding lantibiotic. Additional antibiotic effects, such as induction of autolysis by nisin and Pep5, further complicate the mode-of-action picture, but may well explain why lantibiotics can be so potent and should be stimulating for the construction of new generation antibiotic drugs.

3. Unmodified bacteriocins (class II) Class II bacteriocins are unmodified, cationic and hydrophobic peptides of 20–60 amino acids in length (for a review see [50]). They are divided into three subclass, namely IIa, IIb, and IIc, on the basis of their primary structure. Their activity (in the nanomolar range) mainly induces membrane permeabilisation and leakage of molecules from sensitive bacteria (see Table 1). The inhibition spectrum is rather narrow, limited to species or strains related to the producers. Accordingly, class II bacteriocins are mainly active against low G+C Gram-positive bacteria, such as lactic acid bacteria, Listeria, Enterococcus and Clostridium. 3.1. The subclass IIa bacteriocins Among the so far described class II bacteriocins, more than twenty share high similarities in their primary structure (Fig. 4) as well as an anti-Listeria activity. These bacteriocins are gathered into the subclass IIa, for a recent review

Table 1 Mode of action of class II bacteriocins Mode of action

∆pH

∆ψ

ATP

Efflux

References

Pore formation K+ pores

Dissipation No effect

Dissipation Dissipation

CF K+, CF

[61] [73–75]

Dissipation

Amino acids

[64–70]

Pediocin PA-1

Pore formation PTS ND permease-dependent Pore formation Dissipation

ND Depletion of intracellular ATP ND

Dissipation

Depletion of intracellular ATP

Amino acids K+

[52–60]

Subclass Iib Lactacin F

Pore formation

ND

Dissipation

K+ and Pi efflux

[76]

Lactococcin G

Cation pores

No effect

Dissipation

Amino acids

[78–80]

Plantaricin EF Plantaricin JK Subclass Iic Lactococcin A Lactococcin 972

Cation pores Anion pores

Dissipation Dissipation

Dissipation Dissipation

Depletion of intracellular ATP Depletion of intracellular ATP ND ND

Cations Anions

[77] [77]

Pore formation ND Inhibition of septum ND formation Pore formation Dissipation

Dissipation ND

ND ND

Amino acids No efflux

[82] [83–85]

Dissipation

ND

ND

[81]

Subclass Iia Bavaricin MN Enterocin P Mesentericin Y105

Plantaricin A .

ND, not determined; CF, carboxyfluorescein.

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557

551

Fig. 4. Alignment of amino acid sequences of subclass IIa bacteriocins (by Clustalw). CurA: curvacin A, EntP: enterocin P, BavMN: bavaricin MN, Div41: divercin V41, EntA: enterocin A, Mun: muntcidin, SakP: sakacin P, PedA: pediocin PA-1, MesY: mesentericin Y105, LeuA: leucocin A, CarB2: carnobacteriocin B2.

see [51]. In particular, they bear a highly conserved N-terminal part including the YGNGV consensus motif, a disulfide bridge and several conserved residues. 3.1.1. Pediocin PA-1 Pediocin PA-1 and other identical bacteriocins produced by Pediococcus acidilactici (pediocin AcH, pediocin SJ-1, pediocin JD) are the most extensively studied class II bacteriocins. Bhunia et al. [52] first reported on the mode of action of subclass IIa bacteriocins. They described that treatment with pediocin AcH results in the leakage of K+ and some UV absorbing materials. Pediocin PA-1 was shown to dissipate the membrane potential (∆ψ) of Pediococcus pentosaceus and to cause release of amino acids accumulated either in a proton motive force (PMF)dependent or -independent manner [53]. Furthermore, pediocin PA-1 induces the release of amino acids and other low molecular weight compounds from the vesicles of sensitive cells while it hardly induces carboxyfluoresceine (CF) efflux from liposomes, suggesting that a target protein may be required for bacteriocin activity [53]. Contradictory data have been reported by Chen et al. [54] showing that pediocin PA-1 induces CF efflux in a concentrationdependent manner from liposomes, suggesting that a protein receptor is not required. They also indicate that the binding of pediocin PA-1 to liposomes is dependent on electrostatic interactions and not on the YGNGV consensus motif [55,56]. The presence of a specific target at the cell surface was also suggested by Fimland et al. [57]. In their study, a 15-mer peptide fragment (residue 20–34 from pediocin PA-1) was shown to inhibit the activity of pediocin PA-1 and other subclass IIa bacteriocins (i.e. enterocin A, leucocin A and curvacin A) but to a lesser extent. These data suggest that the 15-mer fragment interferes specifically with the interacting bacteriocins and target cells. Pediocin JD, a bacteriocin that may be identical to pediocin PA-1 [53], was shown to dissipate both the ∆ψ and the ∆pH of Listeria monocytogenes sensitive cells [58]. In addition, Waite et al. demonstrated that pediocin JD might inhibit PEP-mediated glucose uptake via a PTS system [59,60]. It is finally proposed that pediocin PA-1 might modify the permeability of sensitive cells likely by forming pores in

the cytoplasmic membrane and that it needs a specific target molecule at the surface of the sensitive cells. 3.1.2. Bavaricin MN and divercin V41 Bavaricin MN and divercin V41 are two related bacteriocins (Fig. 4) produced by Lactobacillus sake [61] and Carnobacterium divergens, respectively [62]. Bavaricin MN was shown to dissipate both the ∆ψ and the ∆pH of energised L. monocytogenes cells and to induce CF efflux from lipid vesicles in a concentration-dependent manner [61]. Divercin V41 structure–function relationships were studied via chemical modifications and enzymatic hydrolysis [62]. Divercin V41 was cleaved by the endoproteinase Asp-N in two peptides (i.e. 1–17 and 18–43 residues). The second peptide remains active against L. innocua, but the activity was not quantified or compared to that of the whole divercin V41. This result, suggesting that the N-terminal part is not necessary for divercin V41 activity, is contradictory with other data indicating that even minor modifications of the structure of subclass IIa bacteriocins lead to a dramatic drop in the activity. In addition, chemical modifications of divercin V41, leading to net charge variation, were performed to determine their influence on the bacteriocin potency. Indeed, it was proposed that electrostatic interactions between the cationic peptide and the anionic membrane could play a major role in bacteriocin activity. In contrast to the previous postulate, acetylated divercin V41 still retains inhibitory activity, although this derivative is anionic (net charge –2) [62]. The authors suggest that the hydrophobicity of the C-terminal is more important for the activity than its cationic nature. 3.1.3. Mesentericin Y105 and leucocin A Mesentericin Y105 and leucocin A are two related bacteriocins produced by Leuconostoc mesenteroides and Leuconostoc gelidum, respectively. They display a very similar structure with only two differences in their sequence (Fig. 4). The three-dimensional structure of leucocin A was reported [63], a three-strand antiparallel β-sheet domain was found in the N-terminus while the region from position 17 to 31 shows an amphiphilic α-helix conformation. Mesentericin Y105 was shown to exhibit a bactericidal mode of

552

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557

action towards L. monocytogenes [64] and to dissipate the ∆ψ of whole Listeria cells [65]. Even minor modifications either in N- or C-terminus of the mesentericin Y105 sequence impairs the activity and that the disulfide bridge is important for the activity [66], as shown with other subclass IIa bacteriocins. Furthermore mesentericin Y105- and leucocin A-resistant strains were used to investigate the molecular mode of action of these bacteriocin. First, the inactivation of the rpoN gene, encoding the σ54 subunit of the RNA polymerase, results in resistance of L. monocytogenes [67] and E. faecalis [68] to mesentericin Y105. This shows that σ54 directs the expression of a protein responsible for sensitivity. Furthermore, the inactivation of the σ54dependent mpt operon induces resistance of L. monocytogenes [69] and E. faecalis [70]. The mpt operon encodes a mannose permease of the phosphotransferase system (PTS), EIItMan, made of three subunits IIAB, IIC and IID. In addition, increasing glucose or mannose concentration induces simultaneous expression of EIItMan and sensitivity to mesentericin Y105 to L. monocytogenes, suggesting that the expression level of EIItMan affects the sensitivity [69]. Besides, the lack of a mannose PTS subunit expression was first described in a leucocin-resistant L. monocytogenes by Ramnath et al. [71]. This protein is similar to the IIAB subunit of EIItMan, favouring a similar (or identical) mechanism for resistance to mesentericin Y105 and leucocin A. Furthermore, a recent study strongly points out that leucocin A actually requires an interaction with a chiral receptor at the surface of the target cell to be active [72]. Indeed, the D-enantiomer of leucocin A is not active against ten different strains, whereas all of them are sensitive to the natural leucocin A. Taken together these results strongly indicate that EIItMan could be the chiral receptor needed for bacteriocin interaction at the surface of target cells. The membrane-associated IIC and/or IID subunits would interact directly with the bacteriocins. The IID subunit was finally proposed to be the putative target molecule [69]. Alternatively, one could hypothesise that EIItMan is not the target molecule but could control its expression. 3.1.4. Enterocin P Enterocin P is a 44 amino-acid bacteriocin produced by Enterococcus faecium. Enterocin P is secreted by the sec-dependent pathway and exhibits a broad spectrum of activity (i.e. S. aureus, C. perfringens, C. botulinum and L. monocytogenes) compared to most subclass IIa bacteriocins [73]. Enterocin P was shown to dissipate the ∆ψ of previously energised cells but not of non-energised cells [74]. In addition, enterocin P was shown to induce an efflux of 86Rb+ (a K+ analogue) from energised cells, whereas this efflux was not observed from either enterocin P-resistant cells or liposomes [74]. Enterocin P was also shown to deplete intracellular level of ATP from energised sensitive cells even though no appearance of ATP was found in the

external medium [75]. Finally, the authors suggest that enterocin P could form potassium ion-conducting pores in the cytoplasmic membrane of the target cells and that a receptor-like factor might exist in the sensitive cells despite of the rather broad spectrum of antibacterial activity. 3.2. The subclass IIb: two-peptide bacteriocins The subclass IIb includes bacteriocins whose activity depends on the complementary action of two distinct peptides. Accordingly, individual peptides hardly display any activity. There is only one immunity gene, linked to the two structural genes for the peptides, which encodes a unique protein. The primary structure of the two peptides is clearly different and no significant similarity has ever been described between all these bacteriocins (Fig. 5). 3.2.1 . Lactacin F Lactacin F is a two-peptide bacteriocin made from lactacin A (57 amino acids) and LafX (48 amino acids). Lactacin F is produced by Lactobacillus johnsonii and is active against other Lactobacillus and Enterococcus. Lactacin F dissipates ∆ψ, induces a leakage of K+ and decreases the intracellular ATP concentration [76]. This ATP depletion was shown to be likely due to an efflux of inorganic phosphate, resulting in a shift of the ATP hydrolysis equilibrium. These data support that lactacin F likely form pores into the cytoplasmic membrane. 3.2.2. Plantaricin EF and JK Both plantaricin EF and JK are produced by Lactobacillus plantarum C11 and their production is induced by the pheromone/bacteriocin plantaricin A. These two twopeptide bacteriocins were shown to form pores in the target membranes but differ in their ion selectivity. Plantaricin EF dissipates the ∆ψ more efficiently than plantaricin JK and this difference could be due to the higher cation conductance of plantaricin EF [77]. In contrast, plantaricin JK dissipates ∆pH more efficiently than plantaricin EF and was shown to have a higher anion conductance. Plantaricin JK is also more efficient in growth inhibition suggesting that ∆pH is important for cell viability. The authors proposed that a drop in intracellular pH leads to an inhibition of substratelevel phosphorylation, which provides energy to form ATP. 3.2.3. Lactococcin G Lactococcin G is a two-peptide bacteriocin produced by Lactococcus lactis LMG 2081. This bacteriocin is composed of α and β subunits (39 and 35 amino acids, respectively) and is active against several lactic acid bacteria and Clostridium strains [78]. The combination of both peptides is needed to dissipate ∆ψ. Lactococcin G also induces a release of amino acids, such as alanine or leucine, accumulated through a PMF dependent mechanism [79]. In addition, the intracellular ATP level is greatly reduced, which results in a decrease of the ATP-driven glutamate

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557

553

Fig. 5. Primary sequence of subclass IIb and IIc bacteriocins. LafA, LafX: lactacin F; PlnE, PlnF: plantaricin EF, PlnJ, PlnK: plantaricin JK; LcnGα, LcnGβ: lactococcin G; PlnA: plantaricin A; LcnA: lactococcin A; Lcn: lactococcin 972.

uptake. Lactococcin G has no effect on ∆pH suggesting that it does not form proton-conducting pores. In contrast, it leads to a rapid efflux of potassium and other monovalent cations probably due to high selectivity of the pores for these ions [80]. Finally, lactococcin G is inactive against liposomes and membrane vesicles, suggesting that other cell wall-associated components might be involved in sensitivity of target cells. 3.3. The subclass IIc: miscellaneous peptides The subclass IIc and IId were previously defined but no clear structural features support this classification (Fig. 5). In the absence of similarities between several bacteriocins, which could define a new class, we propose that the subclass IIc should include miscellaneous peptides, different from subclass IIa or subclass IIb bacteriocins. 3.3.1. Plantaricin A Plantaricin A is a 26 amino acid peptide synthesised, as well as plantaricin EF and JK, by L. plantarum C11. It exhibits both bactericidal and pheromone activities. A N-terminal truncated form (1–22) was shown to display almost the same activity as the entire peptide. PlnA-22 dissipates both ∆pH and ∆ψ [81]. Its enantiomer, the all-D

PlnA-22 remains bactericidal, whereas it has lost the pheromone activity. This shows that plantaricin A exerts its bactericidal and pheromone activities through different mechanisms and that plantaricin A does not require chiral interactions for its bactericidal activity [81]. 3.3.2. Lactococcin A Lactococcin A is a hydrophobic 54 amino acid peptide produced by L. lactis. It exhibits a very narrow spectrum of activity, limited to other lactococci. Lactococcin A dissipates ∆ψ and induces both the influx and the efflux of amino acids from sensitive cells [82]. It suggests that lactococcin A affects the permeability of the cytoplasmic membrane. In addition, lactococcin A is active against whole cells and membrane vesicles but not against liposomes derived from lactococcal phospholipids [82], suggesting that specific membrane-associated proteins are required for lactococcin A activity. This requirement is in accordance with the very narrow activity spectrum of lactococcin A. 3.3.3. Lactococcin 972 Lactococcin 972 is a 66 amino acid bacteriocin, produced via the sec-dependent pathway by L. lactis IPLA 972 and active against all lactococci tested [83]. Even though the

554

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557

active form of lactococcin 972 was shown to be a homodimer, it could not be classified in the subclass IIb bacteriocin that requires two different peptides. Strikingly, lactococcin 972 is not hydrophobic, as most of the bacteriocins described so far. Another unique feature of lactococcin 972 is that it does not induce any leakage of cytoplasmic solutes. In addition, the effect of lactococcin 972 is not lethal by itself since the target cells keep synthesising macromolecules [84]. Martinez et al. reported that lactococcin 972 inhibits the incorporation of a cell wall precursor, the N-acetylglucosamine, in treated-cells. Lactococcin 972 treatment of sensitive bacteria results in cell elongation and widening likely due to the inhibition of septum formation [85]. This hypothesis is consistent with the observation that lactococcin 972 hardly affect stationary-phase cells.

The bacteriocin first interacts with a target molecule, possibly the EIItMan mannose permease. This interaction is absolutely required for in vivo activity against whole bacterial cells, probably favouring the forthcoming interaction of the bacteriocin with the cytoplasmic membrane. In that case, the structure and the expression level of the target molecule could influence the activity of the bacteriocin. The possible interaction of the bacteriocin with EIItMan could also lead to switch this permease to an open state and could thereby participate to the permeabilisation of the target cell. Secondly, the bacteriocin may interact with the cytoplasmic membrane leading to both pore formation or localised disruption of the membrane. This second step may not require a target molecule and may be mostly dependent on

3.4. Conclusion Subclass IIa bacteriocins tend to dissipate PMF via dissipation of ∆ψ and/or ∆pH (Table 1). It was suggested that bacteriocins themselves could form a pore into the cytoplasmic membrane, although no clear experimental data support this hypothesis. Different models of pore formation were proposed taking into account that these bacteriocins are small amphiphilic peptides [86]. Various factors obviously influence the bacteriocin activity on the bacterial cell. They include the structure and the amount of bacteriocin, the composition and the potential of the cytoplasmic membrane, the structure and the expression level of a protein with an immunity function and the chemical composition of the environment. One of the most important recent findings is that subclass IIa bacteriocins need a target molecule at the surface of sensitive cells to be active. The subclass IIa bacteriocins mainly display a narrow and strain-specific spectrum of activity. The minor differences in phospholipid composition between strains of the same species or between related strains could hardly explain such a high specific spectrum. In addition, the inactivity of all-D leucocin A strongly suggests that a chiral interaction is needed at the bacterial surface for the bacteriocin to be active [72]. This points towards the involvement of a surface protein as a target molecule. Recently, a mannose PTS permease (EIItMan) has been proposed to be a target molecule for mesentericin Y105 [69,70] and leucocin A [71]. Besides, Listeria strains have been shown to be cross-resistant to various subclass IIa bacteriocins while they were still sensitive to nisin, a class I bacteriocin [87] and Y. Héchard (unpublished results). Accordingly, EIItMan is proposed be a target molecule for all subclass IIa bacteriocins in L. monocytogenes, E. faecalis and likely in other sensitive bacteria. On the other hand, several subclass IIa bacteriocins are able to permeabilise the liposomes made of lipids from target cells, supporting the notion that a proteinaceaous target is not necessary for activity. Taking into account these seemingly contradictory results, we can propose the following model for subclass IIa bacteriocin mode of action (Fig. 6).

Fig. 6. A model for the mode of action of subclass IIa bacteriocins. IIAB, IIC and IID represent the subunits of the EIItMan mannose permease (see text). The bacteriocin recognises the permease and then forms pores in the cytoplasmic membrane.

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557

electrostatic and/or hydrophobic interactions with the membrane, explaining why bacteriocins could be active against liposomes or phospholipid vesicles. All subclass IIb bacteriocins studied so far dissipate ∆ψ (Table 1). In addition, lactococcin G dissipates ∆pH while plantaricin EF and JK do not. Another important feature of lactococcin G is that it likely needs a specific receptor molecule to be active. The subclass IIb bacteriocins are proposed to form hydrophilic pores into the membrane of the target cell leading to its permeabilisation, although they could display specific ion selectivity. All these results suggest that subclass IIb bacteriocins exhibit a quite different mechanism of action, in accordance with the differences found among their primary structure. Finally, subclass IIc bacteriocins are miscellaneous peptides with no structure similarity. Accordingly, they are involved in various functions such as membrane permeabilisation, pheromone activity and inhibition of septum formation (Table 1).

Acknowledgements HGS acknowledges support by the DFG (various projects), the European Communities (Contract number QLK2-CT-2000-00411) and by the BONFOR programme of the Medical Faculty, University of Bonn. Anja Hoffmann provided Fig. 1, 2 and 3. YH acknowledges Jacques Frère for providing the Fig. 4, the Rhodia Food company and the ‘Ministère de l’Education Nationale, de la Recherche et de la Technologie’ for their support.

References [1]

[2] [3]

[4]

[5] [6]

[7] [8] [9]

I.F. Nes, D.B. Diep, L.S. Havarstein, M.B. Brurberg, V. Eijsink, H. Holo, Biosynthesis of bacteriocins in lactic acid bacteria, Antonie Van Leeuwenhoek 70 (1996) 113–128. T.R. Klaenhammer, Genetics of bacteriocins produced by lactic acid bacteria, FEMS Microbiol. Rev. 12 (1993) 39–85. H.G. Sahl, R.W. Jack, G. Bierbaum, Biosynthesis and biological activities of lantibiotics with unique post-translational modifications, Eur. J. Biochem. 230 (1995) 827–853. H.G. Sahl, G. Bierbaum, Lantibiotics - Biosynthesis and biological activities of uniquely modified peptides from Gram-positive bacteria, Annu. Rev. Microbiol. 52 (1998) 41–47. A. Guder, I. Wiedemann, H.G. Sahl, Posttranslationally modified bacteriocins–the lantibiotics, Biopolymers 55 (2000) 62–73. G. Jung, Lantibiotics–ribosomally synthesized biologically active polypeptides containing sulfide bridges and didehydroamino acids, Angewandte Chemie, Int. Ed. Engl. 30 (1991) 1051–1068. R.W. Jack, G. Bierbaum, H.-G. Sahl, Lantibiotics and related peptides, Springer-Verlag, Berlin, Heidelberg, New York, 1998. E. Gross, J.L. Morell, The structure of nisin, J. Am. Chem. Soc. 93 (1971) 4634–4635. W. Liu, J.N. Hansen, The antimicrobial effect of a structural variant of subtilin against outgrowing Bacillus cereus T spores and vegetative cells occurs by different mechanisms, Appl. Environ. Microbiol. 59 (1993) 648–651.

555

[10] H.R. Ramseier, Die Wirkung von Nisin auf Clostridium butyricum, Arch. Mikrobiol. 37 (1960) 57–94. [11] P.E. Linnett, J.L. Strominger, Additional antibiotic inhibitors of peptidoglycan synthesis, Antimicrob. Agents Chemother. 4 (1973) 231–236. [12] P. Reisinger, H. Seidel, H. Tschesche, P. Hammes, The effect of nisin on murein synthesis, Arch. Microbiol. 127 (1980) 187–193. [13] E. Ruhr, H.G. Sahl, Mode of action of the peptide antibiotic nisin: influence on the membrane potential of bacterial cells and on cytoplasmic and artificial membrane vesicles, Antimicrob. Agents Chemother. 27 (1985) 841–845. [14] H.G. Sahl, M. Kordel, R. Benz, Voltage-dependent depolarization of bacterial membranes and artificial lipid bilayers by the peptide antibiotic nisin, Arch. Microbiol. 149 (1987) 120–124. [15] H.G. Sahl, H. Brandis, Mode of action of the staphylococcin-like peptide Pep5 and culture conditions effecting its activity, Zbl. Bakt. Hyg. 252 (1982) 166–175. [16] H. Brötz, M. Josten, I. Wiedemann, U. Schneider, F. Götz, G. Bierbaum, et al., Role of lipid-bound peptidoglycan precursors in the formation of pores by nisin, epidermin and other lantibiotics, Mol. Microbiol. 30 (1998) 317–327. [17] E. Breukink, I. Wiedemann, C. van Kraaij, O.P. Kuipers, H.G. Sahl, B. de Kruijff, Use of the cell wall precursor lipid II by a poreforming peptide antibiotic, Science 286 (1999) 2361–2364. [18] I. Wiedemann, E. Breukink, C. van Kraaij, O.P. Kuipers, G. Bierbaum, B. de Kruijff, et al., Specific binding of nisin to the peptidoglycan precursor lipid II combines pore formation and inhibition of cell wall biosynthesis for potent antibiotic activity, J. Biol. Chem. 276 (2001) 1772–1779. [19] G.N. Moll, G.C.K. Roberts, W.N. Konings, A.J.M. Driessen, Mechanism of lantibiotic-induced pore formation, Antonie van Leeuwenhoek 69 (1996) 185–191. [20] G. Bierbaum, H.G. Sahl, Induction of autolysis of staphylococci by the basic peptide antibiotic Pep5 and nisin and their influence on the activity of autolytic enzymes, Arch. Microbiol. 141 (1985) 249–254. [21] G. Bierbaum, H.G. Sahl, Autolytic system of Staphylococcus simulans 22: influence of cationic peptides on activity of N-acetylmuramoyl-L-alanine amidase, J. Bacteriol. 169 (1987) 5452–5458. [22] Z. Oren, Y. Shai, Mode of action of linear amphiphatic, alpha-helical antimicrobial peptides, Biopolymers 47 (1998) 451–463. [23] R. Benz, G. Jung, H.G. Sahl, Mechanism of channel-formation by lantibiotics in black lipid membranes, in: G. Jung, H.-G. Sahl (Eds.), Nisin and Novel Lantibiotics, Escom, Leiden, 1991, pp. 359–372. [24] G. Boheim, Statistical analysis of alamethicin channels in black lipid membranes, J. Membr. Biol. 19 (1974) 277–303. [25] V. Rizzo, S. Stankowski, G. Schwarz, Alamethicin incorporation in lipid bilayers: a thermodynamic study, Biochemistry 19 (1987) 2751–2759. [26] H.W. van den Hooven, C.C.M. Doeland, M. van de Kamp, R.N.H. Konings, C.W. Hilbers, F.J.M. van de Ven, Threedimensional structure of the lantibiotic nisin in the presence of membrane-mimetic micelles of dodecylphosphocholine and of sodium dodecylsulphate, Eur. J. Biochem. 235 (1996) 382–393. [27] H.W. van den Hooven, C.A.E.M. Spronk, M. van de Kamp, R.N.H. Konings, C.W. Hilbers, F.J.M. van de Ven, Surface location and orientation of the lantibiotic nisin bound to membranemimicking micelles of dodecylphophocholine and of sodium dodecylsulphate, Eur. J. Biochem. 235 (1996) 394–403. [28] A.J.M. Driessen, H.W. van den Hooven, W. Kuiper, M. van de Kamp, H.G. Sahl,, R.N.H. Konings, et al., Mechanistic studies of lantibiotic-induced permeabilization of phospholipid vesicles, Biochemistry 34 (1995) 1606–1614.

556

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557

[29] L.Y. Lian, W.C. Chan, S.D. Morley, G.C.K. Roberts, B.W. Bycroft, D. Jackson, Solution structures of nisin and its two major degradation products determined by NMR, Biochem. J. 283 (1991) 413–420. [30] F.J.M van de Ven, H.W. van den Hooven, R.N.H. Konings, C.W. Hilbers, The spatial structure of nisin in aqueous solution, in: G. Jung, H.-G. Sahl (Eds.), Nisin and Novel Lantibiotics, Escom, Leiden, 1991, pp. 35–42. [31] R.A. Demel, T. Peelen, R.J. Siezen, B. de Kruijff, O.P. Kuipers, Z Nisin, mutant nisin Z and lacticin 481 interactions with anionic lipids correlate with antimicrobial activity—a monolayer study, Eur. J. Biochem. 235 (1996) 267–274. [32] E. Breukink, C. van Kraaij, R.A. Demel, R.J. Siezen, O.P. Kuipers, B. de Kruijff, The C-terminal region of nisin is responsible for the initial interaction of nisin with the target membrane, Biochemistry 36 (1997) 6968–6976. [33] C. vanKraaij, E. Breukink, M.A. Noordermeer, R.A. Demel, R.J. Siezen, O.P. Kuipers, et al., Pore formation by nisin involves translocation of its C-terminal part across the membrane, Biochemistry 37 (1998) 16033–16040. [34] C.J. Giffard, S. Ladha, A.R. Mackie, D.C. Clark, D. Sanders, Interaction of nisin with planar lipid bilayers monitored by fluorescence recovery after photobleaching, J. Membr. Biol. 151 (1996) 293–300. [35] K. Winkowski, R.D. Ludescher, T.J. Montville, Physicochemical characterization of the nisin-membrane interaction with liposomes derived from Listeria monocytogenes, Appl. Environ. Microbiol. 62 (1996) 323–327. [36] W.C. Chan, M.L. Leyland, J. Clark, H.M. Dodd, L.Y. Lian, M.J. Gasson, et al., Structure–activity relationships in the peptide antibiotic nisin: antibacterial activity of fragments of nisin, FEBS Lett. 390 (1996) 129–132. [37] V. Arioli, M. Berti, L.G. Silvestri, Gardimycin, a new antibiotic from Actinoplanes, III. Biological properties, J. Antibiot. 29 (1976) 511–515. [38] M. Limbert, D. Isert, N. Klesel, A. Markus, G. Seibert, S. Chatterjee, et al., Chemotherapeutic properties of mersacidin in vitro and in vivo, in: G. Jung, H.-G. Sahl (Eds.), Nisin and Novel Lantibiotics, Escom, Leiden, 1991, pp. 448–456. [39] W.W. Niu, H.C. Neu, Activity of mersacidin, a novel peptide, compared with that of vancomycin, teicoplanin, and daptomycin, Antimicrob. Agents Chemother. 35 (1991) 998–1000. [40] H. Brötz, G. Bierbaum, P.E. Reynolds, H.G. Sahl, The lantibiotic mersacidin inhibits peptidoglycan biosynthesis at the level of transglycosylation, Eur. J. Biochem. 246 (1997) 193–199. [41] H. Brötz, G. Bierbaum, K. Leopold, P.E. Reynolds, H.G. Sahl, The lantibiotic mersacidin inhibits peptidoglycan synthesis by targeting lipid II, Antimicrob. Agents Chemother. 42 (1998) 154–160. [42] N. Zimmermann, J.W. Metzger, G. Jung, The tetracycline lantibiotic actagardine: 1H-NMR and 13C-NMR assignments and revised primary structure, Eur. J. Biochem. 228 (1995) 786–797. [43] E. Racker, C. Riegler, M. Abdel-Ghany, Stimulation of glycolysis by placental polypeptides and inhibition by duramycin, Cancer 44 (1983) 1364–1367. [44] L.L. Chen, P.C. Tai, Effects of antibiotics and other inhibitors on ATP-dependent protein translocation into membrane vesicles, J. Bacteriol. 169 (1987) 2372–2379. [45] J. Navarro, J. Chabot, K. Sherril, R. Aneja, S.A. Zahler, Interaction of duramycin with artificial and natural membranes, Biochemistry 24 (1985) 4645–4650. [46] A. Fredenhagen, F. Märki, G. Fendrich, W. Märki, J. Gruner, J. van Oostrum, et al., Duramycin B and C, two new lanthioninecontaining antibiotics as inhibitors of phospholipase A2, and structural revision of duramycin and cinnamycin, in: G. Jung, H.-G. Sahl (Eds.), Nisin and Novel Lantibiotics, Escom, Leiden, 1991, pp. 131–140.

[47] K. Hosoda, M. Ohya, T. Kohno, T. Maeda, S. Endo, W. Wakamatsu, Structure determination of an immunopotentiator peptide, cinnamycin, complexed with lysophophatidylethanolamine by 1H-NMR, J. Biochem. 119 (1996) 226–230. [48] S.Y. Choung, T. Kobayashi, J. Inoue, Haemolytic activity of a cyclic peptide Ro, Biochim. Biophys. Acta 940 (1988) 171–179. [49] F. Märki, E. Hänni, A. Fredenhagen, J. van Oostrum, Mode of action of the lanthionine-containing peptide antibiotics duramycin, duramycin B and C, and cinnamycin as indirect inhibitors of phospholipase A2, Biochem. Pharmacol. 42 (1991) 2027–2031. [50] I.F. Nes, H. Holo, Class II antimicrobial peptides from lactic acid bacteria, Biopolymers 55 (2000) 50–61. [51] S. Ennahar, T. Sashihara, K. Sonomoto, A. Ishizaki, Class IIa bacteriocins: biosynthesis, structure and activity, FEMS Microbiol. Rev. 24 (2000) 85–106. [52] A.K. Bhunia, M.C. Johnson, B. Ray, N. Kalchayanand, Mode of action of pediocin AcH from Pediococcus acidilactici H on sensitive bacterial strains, J. Appl. Bacteriol. 70 (1991) 25–33. [53] M.L. Chikindas, M.J. Garcia-Garcera, A.J. Driessen, A.M. Ledeboer, J. Nissen-Meyer, I.F. Nes, et al., Pediocin PA-1, a bacteriocin from Pediococcus acidilactici PAC1.0, forms hydrophilic pores in the cytoplasmic membrane of target cells, Appl. Environ. Microbiol. 59 (1993) 3577–3584. [54] Y. Chen, R. Shapira, M. Eisenstein, T.J. Montville, Functional characterization of pediocin PA-1 binding to liposomes in the absence of a protein receptor and its relationship to a predicted tertiary structure, Appl. Environ. Microbiol. 63 (1997) 524–531. [55] Y. Chen, R.D. Ludescher, T.J. Montville, Influence of lipid composition on pediocin PA-1 binding to phospholipid vesicles, Appl. Environ. Microbiol. 64 (1998) 3530–3532. [56] Y. Chen, R.D. Ludescher, T.J. Montville, Electrostatic interactions, but not the YGNGV consensus motif, govern the binding of pediocin PA-1 and its fragments to phospholipid vesicles, Appl. Environ. Microbiol. 63 (1997) 4770–4777. [57] G. Fimland, R. Jack, G. Jung, I.F. Nes, J. Nissen-Meyer, The bactericidal activity of pediocin PA-1 is specifically inhibited by a 15-mer fragment that spans the bacteriocin from the center toward the C terminus, Appl. Environ. Microbiol. 64 (1998) 5057–5060. [58] D.P. Christensen, R.W. Hutkins, Collapse of the proton motive force in Listeria monocytogenes caused by a bacteriocin produced by Pediococcus acidilactici, Appl. Environ. Microbiol. 58 (1992) 3312–3315. [59] B.L. Waite, R.W. Hutkins, Bacteriocins inhibit glucose PEP:PTS activity in Listeria monocytogenes by induced efflux of intracellular metabolites, J. Appl. Microbiol. 85 (1998) 287–292. [60] B.L. Waite, G.R. Siragusa, R.W. Hutkins, Bacteriocin inhibition of two glucose transport systems in Listeria monocytogenes, J. Appl. Microbiol. 84 (1998) 715–721. [61] A.L. Kaiser, T.J. Montville, Purification of the bacteriocin bavaricin MN and characterization of its mode of action against Listeria monocytogenes Scott A cells and lipid vesicles, Appl. Environ. Microbiol. 62 (1996) 4529–4535. [62] P. Bhugaloo-Vial, J.P. Douliez, D. Moll, X. Dousset, P. Boyaval, D. Marion, Delineation of key amino acid side chains and peptide domains for antimicrobial properties of divercin V41, a pediocinlike bacteriocin secreted by Carnobacterium divergens V41, Appl. Environ. Microbiol. 65 (1999) 2895–2900. [63] N.L. Fregeau Gallagher, M. Sailer, W.P. Niemczura, T.T. Nakashima, M.E. Stiles, J.C. Vederas, Three-dimensional structure of leucocin A in trifluoroethanol and dodecylphosphocholine micelles: spatial location of residues critical for biological activity in type IIa bacteriocins from lactic acid bacteria, Biochemistry 36 (1997) 15062–15072. [64] Y. Hechard, C. Jayat, F. Letellier, R. Julien, Y. Cenatiempo, M.H. Ratinaud, On-line visualization of the competitive behavior of antagonistic bacteria, Appl. Environ. Microbiol. 58 (1992) 3784–3786.

Y. Héchard, H.G. Sahl / Biochimie 84 (2002) 545–557 [65] A. Maftah, D. Renault, C. Vignoles, Y. Hechard, P. Bressollier, M.H. Ratinaud, et al., Membrane permeabilization of Listeria monocytogenes and mitochondria by the bacteriocin mesentericin Y105, J. Bacteriol. 175 (1993) 3232–3235. [66] Y. Fleury, M.A. Dayem, J.J. Montagne, E. Chaboisseau, J.P. Le Caer, P. Nicolas, et al., Covalent structure, synthesis, and structurefunction studies of mesentericin Y 105(37), a defensive peptide from Gram-positive bacteria Leuconostoc mesenteroides, J. Biol. Chem. 271 (1996) 14421–14429. [67] D. Robichon, E. Gouin, M. Débarbouillé, P. Cossart, Y. Cenatiempo, Y. Héchard, The rpoN (sigma54) gene from Listeria monocytogenes is involved in resistance to mesentericin Y105, an antibacterial peptide from Leuconostoc mesenteroides, J. Bacteriol. 179 (1997) 7591–7594. [68] K. Dalet, C. Briand, Y. Cenatiempo, Y. Hechard, The rpoN gene of Enterococcus faecalis directs sensitivity to subclass IIa bacteriocins, Curr. Microbiol. 41 (2000) 441–443. [69] K. Dalet, Y. Cenatiempo, P. Cossart, Y. Hechard, A sigma(54)dependent PTS permease of the mannose family is responsible for sensitivity of Listeria monocytogenes to mesentericin Y105, Microbiology 147 (2001) 3263–3269. [70] Y. Hechard, C. Pelletier, Y. Cenatiempo, J. Frere, Analysis of sigma(54)-dependent genes in Enterococcus faecalis: a mannose PTS permease (EII(Man) is involved in sensitivity to a bacteriocin, mesentericin Y105, Microbiology 147 (2001) 1575–1580. [71] M. Ramnath, M. Beukes, K. Tamura, J.W. Hastings, Absence of a putative mannose-specific phosphotransferase system enzyme IIAB component in a leucocin A-resistant strain of Listeria monocytogenes, as shown by two-dimensional sodium dodecyl sulfatepolyacrylamide gel electrophoresis, Appl. Environ. Microbiol. 66 (2000) 3098–3101. [72] L.Z. Yan, A.C. Gibbs, M.E. Stiles, D.S. Wishart, J.C. Vederas, Analogues of bacteriocins: antimicrobial specificity and interactions of leucocin A with its enantiomer, carnobacteriocin B2, and truncated derivatives, J. Med. Chem. 43 (2000) 4579–4581. [73] L.M. Cintas, P. Casaus, L.S. Havarstein, P.E. Hernandez, I.F. Nes, Biochemical and genetic characterization of enterocin P, a novel sec-dependent bacteriocin from Enterococcus faecium P13 with a broad antimicrobial spectrum, Appl. Environ. Microbiol. 63 (1997) 4321–4330. [74] C. Herranz, L.M. Cintas, P.E. Hernandez, G.N. Moll, A.J. Driessen, Enterocin P causes potassium ion efflux from Enterococcus faecium T136 cells, Antimicrob. Agents Chemother. 45 (2001) 901–904. [75] C. Herranz, Y. Chen, H.J. Chung, L.M. Cintas, P.E. Hernandez, T.J. Montville, et al., Enterocin P selectively dissipates the membrane potential of Enterococcus faecium T136, Appl. Environ. Microbiol. 67 (2001) 1689–1692.

557

[76] T. Abee, T.R. Klaenhammer, L. Letellier, Kinetic studies of the action of lactacin F, a bacteriocin produced by Lactobacillus johnsonii that forms poration complexes in the cytoplasmic membrane, Appl. Environ. Microbiol. 60 (1994) 1006–1013. [77] G.N. Moll, E. van den Akker, H.H. Hauge, J. Nissen-Meyer, I.F. Nes, W.N. Konings, et al., Complementary and overlapping selectivity of the two-peptide bacteriocins plantaricin EF and JK, J. Bacteriol. 181 (1999) 4848–4852. [78] J. Nissen-Meyer, H. Holo, L.S. Havarstein, K. Sletten, I.F. Nes, A novel lactococcal bacteriocin whose activity depends on the complementary action of two peptides, J. Bacteriol. 174 (1992) 5686–5692. [79] G. Moll, T. Ubbink-Kok, H. Hildeng-Hauge, J. Nissen-Meyer, I.F. Nes, W.N. Konings, et al., Lactococcin G is a potassium ion-conducting, two-component bacteriocin, J. Bacteriol. 178 (1996) 600–605. [80] G. Moll, H. Hildeng-Hauge, J. Nissen-Meyer, I.F. Nes, W.N. Konings, A.J. Driessen, Mechanistic properties of the two-component bacteriocin lactococcin G, J. Bacteriol. 180 (1998) 96–99. [81] H.H. Hauge, D. Mantzilas, G.N. Moll, W.N. Konings, A.J. Driessen, V.G. Eijsink, et al., Plantaricin A is an amphiphilic alpha-helical bacteriocin-like pheromone which exerts antimicrobial and pheromone activities through different mechanisms, Biochemistry 37 (1998) 16026–16032. [82] M.J. Van Belkum, J. Kok, G. Venema, H. Holo, I.F. Nes, W.N. Konings, et al., The bacteriocin lactococcin A specifically increases permeability of lactococcal cytoplasmic membranes in a voltageindependent, protein- mediated manner, J. Bacteriol. 173 (1991) 7934–7941. [83] B. Martinez, J.E. Suarez, A. Rodriguez, Lactococcin 972: a homodimeric lactococcal bacteriocin whose primary target is not the plasma membrane, Microbiology 142 (1996) 2393–2398. [84] B. Martinez, M. Fernandez, J.E. Suarez, A. Rodriguez, Synthesis of lactococcin 972, a bacteriocin produced by Lactococcus lactis IPLA 972, depends on the expression of a plasmid-encoded bicistronic operon, Microbiology 145 (1999) 3155–3161. [85] B. Martinez, A. Rodriguez, J.E. Suarez, Lactococcin 972, a bacteriocin that inhibits septum formation in lactococci, Microbiology 146 (2000) 949–955. [86] G.N. Moll, W.N. Konings, A.J. Driessen, Bacteriocins: mechanism of membrane insertion and pore formation, Antonie Van Leeuwenhoek 76 (1999) 185–198. [87] M. Rasch, S. Knochel, Variations in tolerance of Listeria monocytogenes to nisin, pediocin PA-1 and bavaricin A, Lett. Appl. Microbiol. 27 (1998) 275–278.